The document analyzes software job advertisements in Turkey to determine the core curriculum for software development programs. It applies association rule mining to 653 job ads and identifies 54 rules among programming languages, frameworks, databases, front-end technologies, and other software skills. The results show that jobs requiring Java, Oracle, and SQL Server experience often require a B.Sc. degree, while C# experience is more likely to position candidates as software specialists. The rules also indicate that C# and SQL Server skills are often required together, while Oracle and Java skills are more specialized.
2. 94 Computer Science & Information Technology (CS & IT)
This increasing importance and rapid change forced the software education to be adaptive to the
market needs. Increasing costs in the software industry, applied wrong strategies, the desires to
rise in quality and performance issues and so fast technology changes had revealed the need to
educate experts in the field and qualified software developers. Therefore, the university-industry
collaboration has gained utmost importance.
Zhengyu stated that a lot of strong software professionals were urgently required in the
community but employers felt that the graduates had the software talent shortage while a
considerable number of graduates could not find a suitable position [3].
In another study, Kuang and Han proposed the methods of teaching reform as guided by market
demand, to update the teaching content, to optimize the teaching methods, to reform the teaching
practice, to strengthen the teacher-student exchange and to promote teachers and students
together because software development training could not meet the needs of the community [4].
Among the software development departments, software Engineering (SE) is the fastest-evolving
engineering discipline that has ability to provide tools and methods for all areas of society [5].
This situation increases the responsibility of SE education to prepare SE professionals for the
industry by providing them with skills to meet the expectations of the software industry.
Innovations and improvements in the curriculum are required to bridge academia-industry gap [6]
since SE education has inability to provide students with large-scale software development
experiences [7]. However, only universities can produce highly skilled professionals who can
satisfy the needs of software industry by taking into account different standards, frameworks and
recommendations developed by interest groups [8].
A study presented software engineering education evolvement in Turkey to provide an
assessment of SE curriculum in Turkish Universities with respect to IEEE/ACM guidelines given
in SEEK (2004) and to provide a guideline to universities conducting an SE programme at
undergraduate level to align their course curriculum with IEEE/ACM guidelines [9].
Students should have necessary background of programming experience for the study of software
engineering concepts in their curriculums. In order to satisfy this condition, the current software
engineering guidelines include concepts and programming paradigms that must be mastered
through study and practice. The well-known guideline for software engineering curricula is
recommended by IEEE/ACM. This guideline gives the standards related to course scheduling,
faculty preparation, student loads, hardware and software resources, instructional materials and
curriculum development. ACM published “Curriculum Guidelines for Undergraduate Degree
Programs in Software Engineering (SE2004)” to provide guidance to academic institutions and
accreditation agencies about what should constitute an undergraduate software engineering
education [10].
This study is related to curriculum development and human resources in software development.
In addition, association rule mining on software job ads is applied. There are similar studies in
literature as follows. Chien and Chen developed a data mining framework for personnel selection
to explore the association rules between personnel characteristics and work behaviors, including
work performance and retention. Moreover, they used decision tree analysis to discover latent
knowledge and extract the rules to assist in personnel selection decisions [11].
3. Computer Science & Information Technology (CS & IT) 95
Mohsin, Ahmad, Din, Mahamud and Din proposed an intelligent model that is aimed at
facilitating key workers select suitable trainees for a training program. In this study, trainees
dataset was mined using association rule to discover important personality characteristics. Their
model produced an efficient selection process and suitable trainees [12].
In another study, Ali and Rajamani presented the solution for selecting appropriate talented
personnel resumes without risk factors using association rule mining. The practical experimental
results obtained from the proposed model encouraged human resource department to take prompt
decisions for recruiting talented personnel accurately without wasting interviewers’ time of
employer and employee. Also, they indicated that the proposed system reduced frequent
resignations, improved performance of talented personnel without training cost and continuous
monitoring [13].
Finally, Smith and Ali indicated that today’s rapid changing and competitive environment
required educators to stay alongside of the job market in order to prepare their students for the
jobs being demanded. They also implied that data mining methods were suitable for this kind of
analysis due to the large volume of job data generated through the web instead of the classical
data analysis methods. Their study illustrated the experience with employing mining techniques
to understand the trend in IT Technology jobs. At the end, collected data from an online agency
was analysed to reach a conclusion about the trends in the job market [14].
In this study, the core part of Software Development Curriculum is determined by applying
association rule mining on Software Job ads in Turkey. As a result, software engineering or
related fields that give education in software development should include these core courses in
their curriculum in order to adapt the software development industry in Turkey.
The rest of this paper is organized as follows: The second chapter makes mention of association
rules and the third chapter depicts how to apply association rule mining on software ads. The
fourth chapter gives results and discussion with respect to the association rule mining. Finally,
the fifth chapter gives the conclusion.
2. ASSOCIATION RULES
One of the important tasks for Knowledge Discovery in data is Association Rule Mining which is
a well-known procedure in data mining. In its basic structure, every association rule fulfilling the
minimum support and confidence are extracted [15]. The general purpose of an association rule A
=> B is to denote that records possessing attribute A also tend to possess attribute B. The aim is
to find association rules which are considered sufficiently interesting as defined by one or more
measures. Most common formulas for support and confidence are as follows [16]:
Support(A=>B)=
ห|∧|ห
|ୈ|
(1)
Confidence(A=>B) =
ห|∧|ห
||||
(2)
where |D| indicates total number of records and ||A|| refers to total number of record including
A.
4. 96 Computer Science & Information Technology (CS & IT)
2.1. Apriori Algorithm
Apriori is an algorithm which is developed for common set learning mining on transactional
database and association rule learning [17]. Apriori uses a level-wise search, where k-itemsets are
used to explore (k+1) itemsets. First, the set of frequent 1-itemsets denoted by L1 is found by
scanning the dataset to find the count for each item, and collecting those items satisfying
minimum support. Then, L1 is used to find L2, the set of frequent 2-itemsets, which is used to
find L3, and so on, until no more frequent k-item sets can be found. The finding of each Lk
requires one full scan of the dataset. The Apriori property is used to improve the efficiency of the
level-wise generation of frequent item sets by reducing the search space [18].
The Apriori property is based on the following interpretations. By definition, if an item set I does
not satisfy the minimum support threshold, min_sup, then I is not frequent (P(I)<min_sup). If an
item A is added to the item set I, then the resulting item set cannot happen more frequently than I.
Therefore, I U A is not frequent either (P(I U A) < min_sup ) [18].
3. ASSOCIATION RULE MINING ON SOFTWARE JOB ADS
3.1 Collecting Data
The data used in this study is taken from one of the popular job recruitment site in Turkey and
currently available ads about software are examined from 5 big cities that are Istanbul (European
Side), Istanbul (Asian Side), Ankara, Izmir and Bursa. At the end feature list and sub-categories
are prepared with respect to job ads for using in the data set [19].
653 software job ads in 5 big cities are collected from this job recruitment site in Excel format.
Finally, 30 main features are determined for finding suitable candidates for the position as shown
in Tab. A.I in Appendix.
3.2 Data Processing
After creating the initial dataset, it is observed that the year of experience in software features can
be important. Consequently, 30 features and some information within these features have been
separated. For example, considering the experience in programming languages on the job
recruitment site, programming languages are divided into 8 pieces as C, C++, C#, Java, Php,
Objective C and the other programming languages. On the other hand, experience levels are
divided into 7 as 0 (Not Acknowledged), 1, 2, 3, 4, 5 and 5+ (Years). Programming language and
experience level features are combined for simplicity. For example, if C# feature contains 2, at
least 2 year-experience is required. In addition, as example qualifications in the special
programming techniques are grouped and new features are created under new names as Software
Architecture Methodologies, Software Patterns, Programming Paradigms and Other Software
Development Processes as shown in Tab. A.II in Appendix. Finally, 653 ads are entered for the
features in Table A.II.
Then, the data is visualized in terms of database systems, programming languages, Front-end
technologies and other software technologies. As shown in Fig. 1, the leading database systems in
5. Computer Science & Information Technology (CS & IT) 97
job ads are MS SQL, Oracle and MySql. The mostly used programming languages are C#, Java
and C++ in job ads as depicted in Fig. 2
Figure 1. The leading database systems in software job ads
Figure 2. The leading programming languages in job ads
In similar way, Fig. 3 shows the leading front-end development technologies in job ads where
Javascript is prominent technology for front-end development. The other software technologies
such as software architecture, software paradigms and web services are essential as shown in Fig.
4.
Figure 3. The leading Front-end technologies in job ads
6. 98 Computer Science & Information Technology (CS & IT)
Figure 4. The leading software technologies in job ads
3.3 Applying Apriori Algorithm
Fig. 1, Fig. 2, Fig. 3 and Fig. 4 above give the frequencies of related technologies but they don’t
give which technologies are related to each other. Moreover, one job ad can contain more than
one software technology at the same time. For this purpose, association rule mining by using
apriori algorithm is applied in assessment. Different combinations trying out through the features
analysis has been done. Minimum support value has been taken as 0.1 and confidence value has
been taken as 0.5 in the analysis process. At the end, 54 rules are determined regarding software
development.
4. RESULTS AND DISCUSSION
The rules obtained by apriori algorithm are divided into 3 parts as follows:
1. Programming languages, frameworks and databases
2. Front-end web technologies
3. Web services, software architectures and programming paradigms
There are 23 rules related to programming languages, frameworks and databases as shown Tab.
1. If the result sides of these rules are noted, Education_Level=3 (B.Sc. Degree), Position=1
(Software Specialist) are prominent results. On the left side of rule 1, {Java=0} condition means
that Java is necessary but the year of experience is not important. The same thing can be said for
C# if the rule 8 is considered. While 87% of ads that want Java experience require at least B.Sc.
degree (Rule 1), 74% of ads that want C# experience require at least B.Sc. degree (Rule 14). It
can be argued from these rules that the university degree is more important in Java than C#.
Looking at rules 6 and 12, the similar result can be obtained for Oracle (79%) and MS SQL
Server (75%). The rules 4 (80%), 5 (80%) show that the ads seeking for C# experience with
MS.NET or ASP.NET framework experience categorize the job seekers as Software Specialists.
In rule 2, the same result can be obtained for 82% of ads seeking for C# experience with MS SQL
Server experience. From the rule 4, 5 and 7, it can be concluded that Job seekers knowing C#
should have enough experience in MS.NET, ASP.NET framework and MS SQL Server. If the
rule 8 is compared with the rule 21, it can be said that while 77% of the ads seeking for C#
7. Computer Science & Information Technology (CS & IT) 99
categorize the job seekers as Software Specialist, the corresponding ratio for Java is 67%. From
these rules, it can be decided that C# experience more valuable for the software firms in Turkey.
The similar result is valid for MS SQL Server (75% in rule 13) and Oracle (71% in rule 18).
Another interesting result can be obtained from the rules 22 and 23 that Oracle ads and Java ads
are separated from each other. However, C# and MS SQL Server are combined in rules 2 and 15.
It means that Oracle and Java are complicated technologies that cannot be known by one
specialist while C# and MS SQL Server are moderate technologies that can be known by one
specialist.
Table 1. Rule Extraction for Programming Languages Frameworks and Database through the Apriori
Algorithm
No Rule Confidence
Value
1 If {Java=0} ==> Education_Level =3 0.87
2 If {C#=0 ∩ MSSQL=0} ==> Position =1 0.82
3 If {Other_Languages =0} ==> Position=1 0.81
4 If {C#=0 ∩ AspDotNetFramework=0} ==> Position =1 0.80
5 If {C#=0 ∩ DotNetFramework=0} ==> Position =1 0.80
6 If {Oracle=0} ==> Education_Level =3 0.79
7 If {AspDotNetFramework =0} ==> C#=0 0.78
8 If {C#=0} ==> Position=1 0.77
9 If {DotNetFramework=0} ==> Position =1 0.76
10 If {DotNetFramework=0} ==> Education_Level =3 0.75
11 If {AspDotNetFramework=0} ==> Position =1 0.75
12 If {MSSQL=0} ==> Education_Level =3 0.75
13 If {MSSQL=0} ==> Position =1 0.75
14 If {C#=0} ==> Education_Level=3 0.74
15 If {C#=0 ∩ MSSQL=0} ==> Education_Level =3 0.73
16 If {C#=0 ∩ DotNetFramework=0} ==> Education_Level =3 0.73
17 If {Other_Languages=0} ==> Education_Level =3 0.72
18 If {Oracle=0} ==> Position =1 0.71
19 If {DotNetFramework =0} ==> C#=0 0.68
20 If {AspDotNetFramework=0} ==> Education_Level =3 0.67
21 If {Java=0} ==> Position =1 0.67
22 If {Java=0} ==> Education_Level =3 ∩ Position =1 0.60
23 If {Oracle=0} ==> Education_Level =3 ∩ Position =1 0.56
20 rules are available about front-end web technologies as shown in Tab. 2. The rule 24 indicates
that 83% of ads that want ajax knowledge require javascript as well (year of experience is not
important). Moreover, 81% of ads that include ajax and software specialist together want
javascript, too. It can be understood from the rules 24 and 25 that ajax and javascript should be
taught together. Similar results can be obtained for the rules 26, 27, 28, 30 in a way that there is a
strong relationship among ajax, jquery, javascript, HTML and CSS in terms of software
development education in front-end web technology. Also, the rules 32, 34, 36, 37 and 39 imply
that at least B.Sc. degree is required at most of the times for front-end development. The rules 24
and 43 show that while 83% of ads that want ajax knowledge require javascript, only 53% of ads
that want javascript require ajax. From these rules, it can be extracted that javascript education is
more fundamental than ajax for front-end education. Another important item for front-end
development is XML as indicated in the rules 32, 33, 40 and 41. The rules 32 and 33 indicate that
8. 100 Computer Science & Information Technology (CS & IT)
the candidates knowing XML should have at least B.Sc. degree and be software specialist at an
important level over 70%. The rules 40 and 41 imply the relation between XML and other web
technologies HTML, CSS and web services. As 62% of ads requiring XML also want HTML and
CSS (rule 40), 56% ads requiring XML want web services. These rules mean that XML is more
common data format for data transfer in web platforms. Actually, a few ads include JSON data
format but they are eliminated by apriori algorithm because of their low support count. It means
that XML is more common in the market in Turkey.
Table 2. Rule Extraction for Front-End Web Technologies through the Apriori Algorithm
No Rule Confidence
Value
24 If {Ajax=0} ==> Javascript=0 0.83
25 If {Ajax=0 ∩ Position =1} ==> Javascript=0 0.81
26 If {Javascript=0} ==> HTML_CSS=0 0.81
27 If {Ajax=0} ==> HTML_CSS=0 0.78
28 If {Jquery=0} ==> HTML_CSS=0 0.78
29 If {Ajax=0} ==> Position =1 0.78
30 If {Jquery=0} ==> Javascript=0 0.76
31 If {Ajax=0 ∩ Javascript=0} ==> Position =1 0.76
32 If {XML=0} ==> Education_Level =3 0.75
33 If {XML=0} ==> Position =1 0.71
34 If {HTML_CSS=0} ==> Education_Level =3 0.70
35 If {Javascript=0} ==> Position =1 0.68
36 If {Ajax=0} ==> Education_Level =3 0.68
37 If {Javascript=0} ==> Education_Level =3 0.67
38 If {Education_Level =3 ∩ Javascript=0} ==> Position =1 0.67
39 If {HTML_CSS=0 Javascript=0} ==> Education_Level =3 0.66
40 If {XML=0} ==> HTML_CSS=0 0.62
41 If {XML=0} ==> Web_Services=0 0.56
42 If {HTML_CSS=0 ∩ Javascript=0} ==> Ajax=0 0.55
43 If {Javascript=0} ==> Ajax=0 0.53
11 rules are obtained related to web services, software architectures and programming paradigms
as depicted in Tab. 3. These rules mostly focus on B.Sc. degree and software specialist position.
In fact, there are other ads focusing on other positions such as database administrator or software
test specialist but they are eliminated because of their low support counts. This means that the
most of software firms give ads focusing on software specialists as shown in rules 49, 51, 52.
From the other perspective, software architectures, web services and programming paradigms
(object oriented programming, functional programming etc.) topics are very special software
technologies that should be carried out by software engineers or equivalents having at least B.Sc.
degree as seen in rules 44, 45, 46, 47 and 48. The rules 53 and 54 imply B.Sc. degree and
software specialist position at the same time. 53% ads including web services imply B.Sc. degree
and software specialist position together (rule 53). The same thing is valid for programming
paradigms (rule 54). The Programming paradigm stands for the styles of various programming
languages such as Phyton, Lisp, F# and Objective-C.
9. Computer Science & Information Technology (CS & IT) 101
Table 3. Rule Extraction for Web Services, Software Architectures and Programming Paradigms through
the Apriori Algorithm
No Rule Confidence
Value
44 If {Software_Architectures=0} ==> Education_Level =3 0.84
45 If {Web_Services=0} ==> Education_Level =3 0.83
46 If {Position=1 ∩ Web_Services=0} ==> Education_Level =3 0.80
47 If {Programming_paradigms=0} ==> Education_Level =3 0.78
48 If {Position=1 ∩ Programming_paradigms=0} ==> Education_Level =3 0.77
49 If {Software_Architectures=0} ==> Position =1 0.68
50 If {Programming_paradigms=0} ==> Position =1 0.68
51 If {Education_Level =3 ∩ Programming_paradigms =0} ==> Position =1 0.67
52 If {Education_Level =3 ∩ Web_Services=0} ==> Position =1 0.64
53 If {Web_Sevices=0} ==> Education_Level =3 ∩ Position =1 0.53
54 If {Programming_paradigms =0} ==> Education_Level =3 ∩ Position =1 0.53
With respect to the rules above, the following courses should be included in Software
Development Curriculum as compulsory core courses as shown in Tab. 4. These courses are
compatible with IEEE/ACM computer science curriculum where it is dictated that successfully
deploying an updated computer science curriculum at any individual institution requires
sensitivity to local needs [20]. The rules above correspond to the local needs in Turkey.
The rules 2, 10, 12 and 15 in Tab. 1 imply that MS SQL Server is a fundamental database system
for software developers. Thus, Database Systems-I course focusing on relational database
concept by MS SQL Server is included in the curriculum. In similar way, since Oracle is also
very popular in the market according to the rules 6, 18, 23 in Tab. 1 Database Systems-II course
focusing on Oracle should be involved. The most popular language in Turkey is C# on MS.Net
framework according to the rules 4, 5, 7, 8, 9, 10, 15, 16 and 19. Because the easiest way to start
to learn C# is Windows Desktop environment, Desktop Programming course should be given in
the curriculum. The rules 1, 21, 22, 47, 48, 50, 51, and 54 in Tab. 1 indicate that Java language is
second popular language in Turkey. Since Java includes all object-oriented principles Object-
Oriented Programming course applying the basic principles on Java is recommended. Web
development can be divided into two parts as front-end and back-end web development. While
the rules 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42 and 43 in Tab. 2
imply Front-End Web Development course, the rules 4, 7, 11, 16, 19 in Tab. 2 involve Back-End
Web Development. Front-end development should include HTML, CSS, Javascript, Jquery,
Ajax, XML. Relating to the rules 4, 7, 11, 16, 19 in Tab. 1, the most popular back-end web
programming is ASP.NET with C# in Turkey. As for the rules 47, 48, 50, 51, 54, the
programming paradigm concept refers to various programming languages having different
programming styles. As a result, Programming Language Concept lecture should be included in
order to teach different programming languages such as Phyton, Lisp, F#, Objective-C. Finally,
Software Architecture course is recommended as compulsory course since the rules 44 and 49 in
Tab. 3 show the importance of software architectures focusing on multi-tier architectures and web
services.
10. 102 Computer Science & Information Technology (CS & IT)
Table 4. The Core Part of Software Development Curriculum for Undergraduate Degree Programs
Year/
Semester
Related Tech. Course Name Related Rules
2/1 MS SQL Server Database Systems-I 7, 12, 13, 15
2/1 C#, MS.Net Framework Desktop Programming 5, 7, 8, 9, 10, 14,
15, 19
2/1 Java Object Oriented
Programming
1, 21, 22, 47, 48,
50, 51, 54
2/2 Oracle Database Systems-II 6, 18, 23
2/2 HTML, CSS, JavaScript,
Jquery, Ajax, XML
Front-End Web
Development
24, 25, 26, 27,
28, 29, 30, 31,
32, 33, 34, 35,
36, 37, 38, 39,
40, 41, 42, 43
3/1 ASP.NET Back-End Web
Development
4, 7, 11, 16, 19
3/1 Phyton, Lisp, F#,
Objective-C
Programming Language
Concepts
47, 48, 50, 51, 54
3/2 Web Services, Multi-Tier Software Architecture 44, 49
5. CONCLUSION
Due to consistent growth in software market and rapid change in software technology, the
adaption of software development curriculum is necessary with respect to criteria desired by
software development industry.
This study has a contribution to the literature in a way that it applies association rule mining on
software job ads to help the managers decide on the software development curriculum. In this
study, the software job ads regarding the first 5 cities having intensive ads are obtained from a
famous Turkish employment web site. While determining the features required for association
rule mining, the years of experience on software expertise areas is considered. After applying
association rule mining, the rules related to desires of software companies are achieved.
Considering the criteria owned by software job ads, it is investigated that practical part of
software development education should be increased. Moreover, since the most of job ads seek
the employee at minimum B.Sc. level, the importance of software engineering or related fields is
increasing as well. As the companies request the new software technologies as well as
fundamental programming abilities, it is vital to revise the software development curriculum at
the universities. However, the most of them are very late to adapt their curriculums to the criteria
of the companies. Actually, it not enough to revise the curriculums but also academic personnel
in these departments should renew their knowledge on the new software technologies.
Since there are many departments related to software development such as software engineering,
computer engineering etc., it is necessary to determine the core courses of software development.
For this purpose, the core courses are determined by using association rule mining on software
job ads and IEEE/ACM computer science curriculum. The core courses should be common for all
departments related to software development because they reflect the local need of the software
development companies in Turkey.
11. Computer Science & Information Technology (CS & IT) 103
As a future study, it is also important to gather the academic personnel and the software company
professionals to focus on the hot software technologies. With respect to the results obtained from
these meetings, the compulsory and elective lectures should be determined so that newly
graduated software developers can easily adapt to the software projects in the market. As a result,
it will be easier to find a job for them and to hire a proper developer for the companies without
giving extra training.
REFERENCES
[1] Laplante, P. A.. What every engineer should know about. Taylor & Francis Group., Boca Raton, FL,
2007 [Online]. Available:http://lib.mdp.ac.id/ebook/Karya%20Umum/Every-Engineer-Should-Know-
about-Software-Engineering.pdf
[2] Mccracken, M., et al. A proposed curriculum for an undergraduate software engineering degree. In:
Software Engineering Education Training. // In Proc. 13th Conference on. IEEE, 2000, p. 246-257.
[3] Zhengyu, G. R. Y. Strengthening Practices and Researches in the Education of Corporation of
Enterprises, Colleges and Institutions. Jiangsu Social Sciences, S2, 2007.
Available: http://en.cnki.com.cn/Article_en/CJFDTOTAL-JHKX2007S2007.htm
[4] Kuang, L. Q.; Han, X. The Research of Software Engineering Curriculum Reform. Physics Procedia.
[Online] 33(2012), pp. 1762-1767.
Available: http://www.sciencedirect.com/science/article/pii/S1875389212015957
[5] Král, J.; Zemlicka, M. Engineering Education-A Great Challenge to Software Engineering. // In
Computer and Information Science, 2008. ICIS 08. Seventh IEEE/ACIS International Conference, pp.
488-495.
[6] Shaw, M.; Herbsleb, J. D.; Ozkaya, I. Deciding What to Design: Closing a Gap in Software
Engineering Education. // Invited paper for Education and Training Track of 27th Int. Conf. on
Software Engineering (ICSE 2005), 2005, pp. 607 – 608.
[7] Su, H.; Jodis, S.; Zhang, H. Providing an integrated software development environment for
undergraduate software engineering courses. // Journal of Computing Sciences in Colleges. [Online]
23, 2(2007), pp. 143-149. Available: http://dl.acm.org/citation.cfm?id=1292453
[8] Jaakkola, H.; Henno, J.; Rudas, I. J. IT Curriculum as a complex emerging process. // In
Computational Cybernetics, 2006. ICCC 2006. IEEE International Conference on 2006, pp. 1-5
[9] Mishra, A.; Yazici, A. An Assessment of the Software Engineering Curriculum in Turkish
Universities: IEEE/ACM Guidelines Perspective. // Hrvatski časopis za odgoj i obrazovanje. [Online]
13, 1(2011), pp. 188-219. Available: http://hrcak.srce.hr/72403
[10] Computing Curricula (CC) (2005). Guidelines for Associate-Degree Transfer Curriculum in Software
Engineering.
Available:http://www.capspace.org/committee/CommitteeFileUploads/TYC_SE_report.pdf
[11] Chien, C. F.; Chen, L. F. Data mining to improve personnel selection and enhance human capital: A
case study in high-technology industry. // Expert Systems with applications. [Online] 34, 1(2008), pp.
280-290. Available: http://www.sciencedirect.com/science/article/pii/S0957417406002776
12. 104 Computer Science & Information Technology (CS & IT)
[12] Mohsin, M. F.; Ahmad, M. F; Din, A. M.; Mahamud, K.R.K.; Din, R. An intelligent trainee selection
model. // In Computer Sciences and Convergence Information Technology (ICCIT), 6th International
Conference on, 2011, pp. 390-
[13] Ali, M. M.; Rajamani, L
manpower considering risk factor: A data mining approach. // In Information Retrieval & Knowledge
Management (CAMP), International Conference on 2012, pp. 39
[14] Smith, D.; Ali, A. Analyzing computer programming job trend using web data mining. // Issues in
Informing Science and Information Technology. [Online] 11, 2014, pp. 203
http://iisit.org/Vol11/IISITv11p203
[15] Ishibuchi, H.; Kuwajima, I.; Nojim
and Pareto-optimality in genetic rule selection. // In Knowledge
Engineering Systems. [Online] 4693, 2007, pp. 509
Available: http://link.springer.c
[16] Agrawal, R.; Imieliński, T.; Swami, A. Mining association rules between sets of items in large
databases. // In ACM SIGMOD Record. 22, 2(1993), pp. 207
[17] Agrawal, R.; Srikant, R. Fast algorithms for mi
large data bases. VLDB 1215, 1994, pp. 487
[18] Han, J.; Kamber, M. Data Mining, Southeast Asia Edition: Concepts and Techniques. Morgan
Kaufmann. 2006.
[19] Job Recruitment Website (JRW). 2
[20] Curriculum Guidelines for Undergraduate Degree Programs in Computer Science (CGCS) (2013,
December). Computer Science Curricula 2013 [Online].
Available: http://www.acm.org/education/CS2013
AUTHORS
Ilkay Yelmen received the BS degree in Software Engineering from the Istanbul
Aydin University in 2013 and he is currently pursuing the MS degree in Computer
Engineering at the Istanbul Technical University. His research interests are Data
Mining and Natural Language Processing.
Metin Zontul received the BS Degree in Computer Engineering from Middle East
Technical University, Ankara, Turkey, the MS Degree in Computer Science from
Erciyes University, Kayseri, Turkey, and the PhD Degree in
from Cumhuriyet University, Sivas, Turkey. His research interests are software
development, information systems, soft computing and data mining. He is currently
a faculty member of Software Engineering Department at Istanbul Aydin Univers
Istanbul, Turkey.
Computer Science & Information Technology (CS & IT)
in, M. F.; Ahmad, M. F; Din, A. M.; Mahamud, K.R.K.; Din, R. An intelligent trainee selection
model. // In Computer Sciences and Convergence Information Technology (ICCIT), 6th International
-393.
Ali, M. M.; Rajamani, L. Automation of decision making process for selection of talented
manpower considering risk factor: A data mining approach. // In Information Retrieval & Knowledge
Management (CAMP), International Conference on 2012, pp. 39-44.
lyzing computer programming job trend using web data mining. // Issues in
Informing Science and Information Technology. [Online] 11, 2014, pp. 203-214. Available:
http://iisit.org/Vol11/IISITv11p203-214Smith0494.pdf
Ishibuchi, H.; Kuwajima, I.; Nojima, Y. Prescreening of candidate rules using association rule mining
optimality in genetic rule selection. // In Knowledge-Based Intelligent Information and
Engineering Systems. [Online] 4693, 2007, pp. 509-516.
Available: http://link.springer.com/chapter/10.1007/978-3-540-74827-4_64
ski, T.; Swami, A. Mining association rules between sets of items in large
databases. // In ACM SIGMOD Record. 22, 2(1993), pp. 207-216.
Agrawal, R.; Srikant, R. Fast algorithms for mining association rules. // In Proc. 20th int. conf. very
large data bases. VLDB 1215, 1994, pp. 487-499.
Han, J.; Kamber, M. Data Mining, Southeast Asia Edition: Concepts and Techniques. Morgan
Job Recruitment Website (JRW). 2014. Available: http://www.kariyer.net/
Curriculum Guidelines for Undergraduate Degree Programs in Computer Science (CGCS) (2013,
December). Computer Science Curricula 2013 [Online].
Available: http://www.acm.org/education/CS2013-final-report.pdf
Ilkay Yelmen received the BS degree in Software Engineering from the Istanbul
Aydin University in 2013 and he is currently pursuing the MS degree in Computer
Engineering at the Istanbul Technical University. His research interests are Data
Mining and Natural Language Processing.
Metin Zontul received the BS Degree in Computer Engineering from Middle East
Technical University, Ankara, Turkey, the MS Degree in Computer Science from
Erciyes University, Kayseri, Turkey, and the PhD Degree in Numerical Methods
from Cumhuriyet University, Sivas, Turkey. His research interests are software
development, information systems, soft computing and data mining. He is currently
a faculty member of Software Engineering Department at Istanbul Aydin University,
in, M. F.; Ahmad, M. F; Din, A. M.; Mahamud, K.R.K.; Din, R. An intelligent trainee selection
model. // In Computer Sciences and Convergence Information Technology (ICCIT), 6th International
. Automation of decision making process for selection of talented
manpower considering risk factor: A data mining approach. // In Information Retrieval & Knowledge
lyzing computer programming job trend using web data mining. // Issues in
214. Available:
a, Y. Prescreening of candidate rules using association rule mining
Based Intelligent Information and
ski, T.; Swami, A. Mining association rules between sets of items in large
ning association rules. // In Proc. 20th int. conf. very
Han, J.; Kamber, M. Data Mining, Southeast Asia Edition: Concepts and Techniques. Morgan
Curriculum Guidelines for Undergraduate Degree Programs in Computer Science (CGCS) (2013,
13. Computer Science & Information Technology (CS & IT) 105
Appendix
Table AI. Main Features and Their Sub Details in Software Job Ads
14. 106 Computer Science & Information Technology (CS & IT)
Table AII. The Final Dataset Structure for Software Job Ads